The two-point Padé approximation problem and its Hankel vector (Q6569148)

In two-point Padé approximation one has to find two coprime polynomials such that the power series expansion of their ratio matches as far as possible the initial terms of two power series of a function \(f\), one at 0 and one at \(\infty\). The linearized problem is obtained by multiplying the defining interpolation conditions with the denominator polynomial. The two power series of \(f\) are combined such that it results in an auxiliary classical Padé problem at \(\infty\). The particularity of this paper is that the expansion at \(\infty\) is allowed to have a polynomial part \(d\). That is solved by considering the problem for \(g=f-d\) and adding \(d\) to its solution. The vector of coefficients that are matched at \(\infty\) in the modified Padé problem is called the Hankel vector. The rank structure of the Hankel matrix whose entries are the Hankel vector defines the solvability of the problem and the recurrence for the numerator and denominator polynomials. Hence, this also allows some properties about the block structure in the two-point Padé table.